A Comparison of Basic Patterns
Stephen J. Moody
A conversation consisting of neutral statements of known fact is likely to be a boring one. Much intrigue in human communication lies in examining evidence and using factual information as a basis for making interesting and often insightful inferences. This idea of making a conjecture based on some form of evidence, hereafter referred to as “evidentiality”, is an intriguing study and can teach us a great deal about language patterns and the way a culture judges perceived information. Roy Freedle stated the importance of evidentiality well.
[Evidentials] can, for one thing, show us much about what we might regard as ‘natural epistemology’, the ways in which ordinary people, unhampered by philosophical traditions, naturally regard the source and reliability of their knowledge. Simultaneously we can learn a great deal about an important ingredient of language itself, the ways in which languages agree and differ in their emphases, and in the kinds of devices they make available to their speakers (vii).
We are familiar with evidentiality in English by phrases such as “it looks like”, “it seems as if”, “I heard”, “he says”, “it appears that”, and “it resembles”. There are dozens of similar grammar patterns that can be included in this category. Of course, evidentiality is not unique to English and can be found in most forms of communication#. This paper seeks to describe a few of the basic evidentials unique to the Japanese language, how they are formed into grammatically correct sentences, and what different nuances are associated with each one.
Four specific evidentials will be studied in separate sections including yō, mitai, rashii, and sō. It is important to emphasis that these four constructions by no means compose an exhaustive list. They are intended merely to highlight the most common patterns to help the reader develop a general feel for Japanese evidentiality. The conclusions presented can be used to develop an understanding of how the Japanese make a conjecture based on observed information. We will examine a few of the disagreements concerning the grammatical classification of evidential particles. The paper will also argue that, when forming evidential patterns, there are exceptions to standard grammar rules as they might appear in introductory texts. We will see that, while there are many circumstances when different patterns mean the same thing, often it is the context which dictates when one construction is preferred over another.
An analysis will also be developed based on a theory presented by Akio Kamio, which he refers to as “territory of information” (see Kamio 1). This is the idea that certain grammar patterns are used based on the source of information, whether or not the speaker posses it, and what it refers to. It attempts to analyze information used when making an utterance and classify it as either inside or outside of the either the speaker’s or the listener’s territory of information. Kamio explains that “information falling into the speaker’s or hearer’s territory of information may be quite informally characterized as information that the speaker/hearer considers proximal, or ‘close’, to him/herself” (2).
As the subtleties of a language are difficult, perhaps impossible, to understand perfectly by memorizing a set of rules, it is hoped that the reader will be able to gain insight into the basics of Japanese evidentiality and then draw his or her own conclusions. This will allow the reader to gain personal insight and develop a deeper understanding of not only the Japanese language, but also its culture and the methods of communication utilized by native speakers. To this end, the paper seeks not to perfectly describe the rules and rigor of syntax and semantics, but rather to establish a comparison between several basic forms and provide a framework for determining what situations call for what patterns.
Among the most basic evidentials in Japanese are constructions the particle yō. An attempt to translate this word directly into English typically results in something similar to ‘seems’, ‘looks like’, or ‘appears’. Although these definitions get close to the actual meaning well enough to help a beginning student learn the language, a more advanced student will realize that achieving a concise translation is a much more involved process. We first look at how yō appears in typical Japanese sentences and then use this as a base to examine the precise meaning as it may be understood by a native speaker.
As a logical first step in determining what role yō plays in a Japanese sentence it is natural to first define its part-of-speech. This is, however, a slightly controversial task.# The difficulty in classifying it comes from the fact that it really does not behave like any of the standard Japanese parts-of-speech. We will see that it occasionally makes sense to treat it like a meishi while other times it is behaves like a keiyōdōshi. This section argues that it is most appropriate to classify it as a jodōshi, but admit that functionally it tends to behave like either a meishi and or a keiyōdōshi.
The first thing that all students of Japanese learn is that yō must always be preceded by a phrase and cannot stand alone. It follows that many basic grammar books present it as an auxiliary particle, or something similar (see Makino and Tsutsui 547-64). However, an argument is often made should be treated like a meishi. Indeed, it would appear to function in this manner in many situations. This is demonstrated most effectively with a few simple examples:
1a) Jon wa Tōkyō ni iku yō desu.
It seems that John will go to Tokyo.
1b) Nihongo wa amerikajin niwa muzukashii yō desu.
It seems that Japanese is difficult for Americans.
1c) Aiko wa Takeru no koto ga suki na yō desu.
It seems that Aiko likes Takeru.
All of the modifying words in these three sentences, classified as dōshi (as in sentence 1a), keiyōshi (1b), and keiyōdōshi (1c), respectively, connect to yō using the rentaikei conjugation. This is how we would expect them to modify any other noun and it seems to be the basis from which linguistics may argue that yō is indeed a meishi.
However, another argument asserts that yō is really a keiyōdōshi. This is easily understood when one considers a scenario when yō itself modifies another word.
1d) Sumisu-san wa nihonjin no yōni hanashite imasu.
Mr. Smith is speaking Japanese like (appearing is if he is) a Japanese.
1e) Ano hito wa saru no yōna kao wo shite iru.
That person has a face resembling a monkey.
Thus yō appears to modify verbs by adding ni and nouns by adding na, which is precisely how a keiyōdōshi modifies other words (Makino and Tsutsui 547). In summary, to construct a sentence using yō, it is modified by attaching a word or phrase to the left side as if it is a meishi, but when it modifies other words on the right side, it is conjugated as if it is a keiyōdōshi.
The above argument admits there are times when yō behaves as several different parts of speech and this is probably why there is a lack of harmony amongst grammarians considering it’s appropriate classification. It is easy to see that it really cannot be satisfactorily labeled as one or the other. With this in mind, note that because yō cannot begin a sentence or stand by itself, it violates the rules of being jiritsugo altogether. So really the only part of speech left is jodōshi, and indeed this classification seems to be the favorite of most authors (see McClain 1, 80).
Martin presents a very intriguing analysis. He points out the subtle fact that yō may actually nominalize the phrase preceding it (911). Consider the following:
1f) Mae ni hanashita yōni ashita mise ni ikimasu.
As I told you before (in the manner in which I previously told you), I will go to the store tomorrow.
Thus the phrase mae ni hanashita ‘[I] told you before’ is nominalized by attaching it to yō. This is made more obvious by comparing it to a standard nominalizer, such as koto, in which case the phrase would become mae ni hanashita koto ‘the thing of telling you previously’. This is a neutral nominalization wherein what was originally a verb phrase is now treated in its entirety as a noun. By this same principle, mae ni hanashita yō is also a nominalized phrase. The only difference is the conjectural nature of the latter. Thus it could be translated as ‘the manner in which I told you before’. It can now be used as a modifier of other words, functioning precisely as if it is a keiyōdōshi#. In this sense, an analogy can be constructed in which koto turns phrases into meishi and yō turns phrases into keiyōdōshi, which is indeed an interesting idea.
Martin has broken down yō into three broad definitions, ‘as if S is so’, ‘in a way so that S is so’, and ‘in a way which agrees with S’ (741). This is perhaps the most standard definition. Yō indicates resemblance and is used to compare an agent to nouns, verbs, or adjectives. In this sense, it is a contrastive particle used in making comparisons by suggesting that one thing looks or behaves like something else (Vaccari 404).
Makino and Tsutsui also make an interesting argument by saying that yō is used for conjecture based on reliable first-hand information from which the speaker then makes some inference. When the speaker indicates that it seems as if John went to Tokyo, as in example 1a, the use of yō indicates the speaker feels confident this conjecture is accurate. Thus yō carries with it a high degree of certainty (552).
The territory of information principle can also lend a hand toward understanding the nuances of yō. Yō is based on reliable first-hand evidence and is thus information “owned” by the speaker in the sense that he has obtained it and understands what it is. This information is then used to make a logically based conjecture about something the speaker may not know. In English this may be expressed by saying, “Based on the observed facts, I conclude that it appears . . . .” This idea will become important in the following sections when we compare yō to other evidentials.
Having established the construction and meaning of yō, one of the most basic evidential forms in Japanese, this paper will now look at three other constructs which have comparable functions. It will describe them using a similar framework, but with references to the ideas just presented. There will be an occasional attempt to explicitly compare and contrast the various forms, but it should be noted that the ultimate decision on when to use a specific grammar patter is left to the reader’s discretion upon gaining an understanding of the nuances conveyed.
Occasionally considered the colloquial equivalent of yō#, mitai is the logical next-step in a development of evidential constructions in Japanese. As will be explained, mitai is indeed used in ways resembling that of yō. However, an argument will also be presented suggesting that such a claim is not very accurate. From a purely syntactical standpoint, mitai is analogous to yō, but the analogy is far from perfect. Mitai clearly differs in the nuances portrayed to the listener and there are situations when one is undoubtedly preferred over the other.
For reasons analogous to yō, mitai is sometimes presented as a keiyōdōshi and sometimes as a meishi. When mitai itself is used to modify another word, the former is valid. Consider the following:
2a) Sushi mitai na tabemono ga kirai desu.
(I) dislike food that resembles sushi.
2b) Tsuyu mitai ni ame ga yoku furimasu ne.
It rains often as if it is the rainy season.
2c) Ame ga furu mitai desu.